Multinational conglomerate based in China, Alibaba Group Holding, is preparing to unveil an enhanced version of its flagship artificial intelligence (AI) model, Qwen 3, later this month, according to reports from Bloomberg. The exact timing of the release may still shift, as indicated by an unnamed source.
Qwen 3 is an upgraded version of AI model, part of the Qwen series. This model is designed to enhance various AI capabilities, including natural language processing, image recognition, and multimodal input processing, which allows it to handle text, images, audio, and video.
The Qwen models aim to provide advanced generative AI functionalities, enabling applications in areas like content creation, customer service, among others.
While Alibaba has not yet responded to requests for comment, the upcoming launch comes amid increasing competition in the AI landscape. Chinese firms have been rolling out affordable AI services, particularly following the attention garnered by DeepSeek, which claimed its model was cheaper to develop compared to those from American competitors.
In recent weeks, companies like OpenAI, Google, and Anthropic have also introduced new AI models, further intensifying the competitive atmosphere.
Earlier this year, Alibaba released various AI products, including a new model in the Qwen 2.5 series last week and an updated version of its AI assistant, Quark, just last month. The company has announced plans to significantly boost its AI investment over the next three years, surpassing what it has spent in the past decade.
During a February earnings call, CEO Eddie Wu emphasized the company’s commitment to achieving artificial general intelligence (AGI). “We aim to continue developing models that push the boundaries of intelligence,” he stated. “Our focus on expanding these boundaries is driven by the need to create new opportunities across various AI applications.”
Recently, Alibaba Cloud introduced the Qwen2.5-Omni-7B model, which is designed to handle multiple input types—text, images, audio, and video—and generate real-time responses in text and natural speech. The company highlighted that this compact, multimodal model enables the creation of agile and cost-effective AI agents.
